Hackernews
new
show
ask
jobs
XPress: Parallel Refinement for Diffusion Drafters in Speculative Decoding
2 points
posted 10 hours ago
by matt_d
(supercomputing-system-ai-lab.github.io)
No comments yet